2010 DOE Hydrogen Program

PI: F. Colin BusbyW. L. Gore & Associates, Inc.

6/11/2010 Project ID #MN004

This presentation does not contain any proprietary, confidential, or otherwise restricted information.

MANUFACTURING OF LOW-COST, DURABLE MEMBRANE ELECTRODE ASSEMBLIES

ENGINEERED FOR RAPID CONDITIONING

• Project start: October 1, 2008• Project end: March 31, 2012• 35% Percent Complete as of 4/8/10

• Total Project Funding: $4.4MM– $2.8MM DOE Share– $1.6MM Contractor Share

• Received in FY09: $654k– $890k spent as of 4/8/10

• Funding for FY10: $700k

TimelineBudget

Barriers Addressed

Partners

Overview

• Lack of High-Volume MEA Processes

• Stack Material & Mfg. Cost

• MEA Durability

• University of Delaware (UD) – MEA Mechanical Modeling

• Penn State University (PSU) – Fuel Cell Heat / Water Management

Modeling & Validation• UTC Power, Inc. (UTCP)

– Stack Testing• W. L. Gore & Associates, Inc. (Gore)

– Project Lead

Relevance: Overall ObjectiveThe overall objective of this project is to develop unique, high-volume1

manufacturing processes that will produce low-cost2, durable3, high-power density4 3L MEAs5 that require little or no stack conditioning6.

1. Mfg. process scalable to fuel cell industry MEA volumes of at least 500k systems/year2. Mfg. process consistent with achieving $15/kWe DOE 2015 transportation stack cost target3. The product made in the manufacturing process should be at least as durable as the MEA

made in the current process for relevant automotive duty cycling test protocols4. The product developed using the new process must demonstrate power density greater or

equal to that of the MEA made by the current process for relevant automotive operating conditions

5. Product form is 3 layer MEA roll-good (Anode Electrode + Membrane + Cathode Electrode) 6. The stack break-in time should be reduced by at least 50 % compared to the product made

in today’s process, and break-in strategies employed must be consistent with cost targets

Relevance: Objectives• Develop a relevant high-volume cost model for membrane

electrode assembly (MEA) production and estimate potential savings for an MEA made in an improved process (Gore)– Estimate potential savings to justify kick-off of new process

development FY ‘09 – Evaluate potential for new process to achieve DOE cost

targets prior to process scale-up ( Go / No-Go Decision) FY ‘11

• R&D equipment procurement and qualification (Gore)– Low-cost membrane coating FY ‘09– Low-cost cathode electrode coating FY ‘09– Low-cost anode electrode coating FY ‘09

Relevance: Objectives• Low-cost MEA R&D

– New Process Exploration (Gore) FY ’09 & ‘10• Investigate equipment configuration for MEA production • Investigate raw material formulations • Map out process windows for each layer of the MEA

– Fuel Cell Heat & Water Management Modeling & Validation (PSU) FY ’10• Low cost processes result in a different range of electrode properties• Electrode and GDL thermal, geometric, & transport properties and

interactions need to be optimized efficiently to meet project goals– Multi-layer Mechanical Modeling (UD) FY ’09 & ’10

• Develop a deeper understanding of MEA failure mechanisms• Use model to optimize mechanical durability of the MEA

structure targeted by the new low-cost process– MEA optimization FY’ 11

• Utilize modeling results & designed experiments

• Conditioning FY ‘11• Scale Up FY ‘11• Stack Validation FY ’11

Approach: Summary• Reduce MEA & Stack Costs

– Reduce the cost of intermediate backer materials which are scrapped

– Reduce number & cost of coating passes – Improve safety & reduce process cost by

minimizing use of solvents – Reduce required conditioning time & costs

• Optimize Durability – Balance tradeoffs between mechanical properties

and performance of the 3L construction

• Enabling Technologies: – Direct coating: Use coating to form at least

one membrane–electrode interface– Gore’s advanced ePTFE membrane

reinforcement & advanced PFSA ionomers enable durable, high-performance MEAs

– Utilize modeling of mechanical stress and heat / water management to accelerate low-cost MEA optimization

– Advanced fuel cell testing & diagnostics

• Low-cost MEA R&D– R&D Equipment Procurement and Qualification– Primary path

• Direct coated cathode on a backer-supported reinforced membrane

• Direct coated anode on 2-layer intermediate– Alternate path

• Direct coat anode on supported ½ membrane• Direct coat cathode on supported ½ membrane• Bond the membrane-membrane interface of the

1.5 layer webs to make a 3-layer web

Approach:

Approach: Low-Cost MEA Mfg Process: Primary Path

High-Volume, Low-Cost, Full Width MEA Production Reduce the cost of intermediate backer materials which are scrapped Reduce number & cost of coating passes Improve safety & reduce process cost by minimizing use of solvents Reduce required conditioning time & costs

Electrode Ink

Membrane on support

2L MEA Intermediate

Oven

Oven

Electrode Ink

2L MEA Intermediate

Oven

Oven 3L MEA Final Product

Support take-up

Approach: Mechanical Modeling• Model Concept:

Develop a layered structure MEA mechanical model using non-linear (viscoelastic & viscoplastic) membrane and electrode properties to predict MEA stresses for input temperature & relative humidity cycling scenarios

• Devise & perform experiments to determine mechanical properties of MEA materials as functions of:

– Temperature– Humidity – Time

• Use numerical modeling to predict mechanical response of MEA during cell operation and accelerated testing

• Use model to explore new MEA designs and optimize prototypes that will be made in the new low-cost process

0

5

10

15

20

25

30

35

0 0.2 0.4 0.6 0.8 1 1.2

True

Stre

ss (M

Pa)

True Strain (mm/mm)

T=25˚C

T=45˚C

Nafion 211In Water

T=65˚C

T=85˚C

MTS Tensile Tester RT/5Approach: Mechanical Modeling

MTS Alliance RT/5 Material Testing System & ESPEC Environmental Chamber

Humidity controlled environment

Temperature and Relative Humidity Capability: T: 25 ˚C - 85 ˚C; RH: 30-90% 0

10

20

30

40

50

60

70

80

0 0.2 0.4 0.6 0.8 1 1.2 1.4

100 mm/min

10 mm/min

1 mm/min

0.5 mm/min

0.1 mm/min

0.05 mm/min

0.01 mm/min

True Strain (mm/mm)

True

Stre

ss(M

Pa)

T=25˚C RH=30%

NAFION is a registered trademark of E. I. DuPont de Nemours & Company

NAFION® 211In Water

Constitutive Model: Visco-elastic-plastic Model

Dashpot ElementStrain-rate dependence

Viscous power law

εησ =

nvv A )(σε =

Spring ElementStrain dependence

(Long-term modulus)

(Instantaneous modulus)

εσ K=

vP KK +

PK

Elastoplastic terms

Visco-Elastic terms

Elastic Plastic

Elastic

( )vP

v

KKKf+

=

Parameters

HKKE

fnA

yield

VP

,,),(

,,,,

συ

λθ+

Viscous

Approach: Mechanical Modeling

Potential Model Output Data:• Water Transport and MEA Stresses

- 3L Residual In-plane Stress After De-Hydration

Technical Accomplishments & Progress: Summary

• Test Current Commercial MEA (Gore)– Power density baseline testing FY08 Completed– Conditioning baseline testing FY08 Completed– Mechanical durability baseline testing FY08 Completed– Chemical durability baseline testing FY08 Completed

• Cost Model Current Commercial MEA (Gore)– Model generic decal lamination process FY08 Completed– Perform raw material sensitivity analysis 100% Complete

• Mechanical Modeling (UD)– Layered model development 90% Complete– RH & time-dependent mechanical testing 35% Complete

Technical Accomplishments & Progress: Summary

• Equipment Procurement and Qualification (Gore) – Membrane coating 100% Complete– Cathode coating 100% Complete– Anode coating 60% Complete

• MEA Alternative Concepts Generation and New Process Design (Gore)

– Process feasibility screening 75% Complete

– Determine primary and alternative paths 100% Complete– Direct coated cathode 25% Complete

• Power density baseline testing• Electrochemical diagnostics

– Direct coated anode 35% Complete• Power density baseline testing

Technical Accomplishments:3L MEA Manufacturing Process Cost Model

Cost model results indicate that a new 3L MEA process has potential to reduce MEA cost by 25%

Process Waste Map

Sensitivity Analysis

0 50 100 150 200 250 300 350 400

Pt reclaim credit from manufacturing scrap

Ionomer cost ($ / lb)

Membrane process cost ($ / m2)

Combined electrode coating process cost ($ / m2)

Catalyst process mark-up

Pt cost ($ / Troy Ounce)

Total loading (mg Pt-Group Metal / cm2)

3L Cost ($/m2)

HighBase CaseLow

Technical Accomplishments: Single-Variable Sensitivity Analysis Base Case (0.4 mg Pt/cm2 total loading), Improved Process

References1 2008 TIAX Tech Team Review "Direct Hydrogen PEMFC Manufacturing Cost Estimation for Automotive Applications" DE-AD36-06GO260442 2005 TIAX Report "Cost Analysis of PEM Fuel Cell Systems for Transportation" NREL/SR-560-391043 2004 TIAX "Platinum Availability and Economics for PEMFC Commercialization" DE-FC04-01AL676014 http://www.platinum.matthey.com/prices/price_charts.html5 http://www1.eere.energy.gov/hydrogenandfuelcells/mypp/pdfs/fuel_cells.pdf

Category Description Base Case Sensitivity Range Sensitivity notes Pt cost ($ / Troy Ounce) 1100 640 - 2060 Min: mean real price3 Max: 2008 monthly avg maximum4

Catalyst process mark-up 20% 5 - 30% Based on Gore interactions with catalyst suppliersTotal loading (mg Pt-Group Metal / cm2) 0.40 0.2 - 0.75 Min: DOE 2015 target5, table 3.4.12 Max: 2005 TIAX Report2 Table 14Pt reclaim credit from manufacturing scrap 95% 90 - 98% Based on Gore interactions with catalyst suppliers

Ionomer Ionomer cost ($ / lb) 88 22 - 110 2005 TIAX Report2 Table 20 ($20 - $100/lb in 2005 dollars)Membrane process cost ($ / m2) 2.18 1.00 - 4.00 Gore assumption, MOH & DL onlyCombined electrode coating process cost ($ / m2) 6.00 3.00 - 12.00 Gore assumption, MOH & DL only

Catalyst

Coating

Key Input Assumptions and Justification

, adapted from TIAX methodology, adapted from TIAX methodology

Technical Accomplishments:Direct Coated (DC) Electrodes

Scanning electron microscopy images of direct-coated electrodes on GORE-SELECT® Membrane

Membrane

DC CathodeMembrane

DC Anode

GORE, GORE-SELECT and designs are trademarks of W. L. Gore & Associates, Inc.

Technical Accomplishments:DC Anode: Power Density

Performance of an MEA made with DC anode is comparable to a commercial control MEA. Due to process challenges that have since been overcome, DC anode Pt loadings were higher than the current commercial control anode for this data set.

Wet Very Wet

Dry Very Dry

Technical Accomplishments:Direct Coated Cathode: Power Density

DC Cathode (red) mass transport

performance is low in super-saturated

operating conditions

DC Cathode performance (red) is comparable to

commercial control MEA (blue)

Wet Very Wet

Very DryDry

Technical Accomplishments:Direct Coated Cathode: Power Density

• Current interrupt resistance accounts for most of the difference between the direct coated cathode MEA and the commercial control MEA in the ohmic region of the pol curve (~200 to ~1,000 mA/cm2)

CIR Adjusted

Standard Pol Curves

Direct Coated

Control

Very Dry Very Dry

Technical Accomplishments:DC Cathode Electrochemical Diagnostics

• Current interrupt resistance and high frequency resistance correlate well with performance data in FA conditions, where ohmic losses are very significant

• Electrode ionic resistance is inversely correlated to performance • Direct coating changes water management – modeling has potential to accelerate development

Very Dry

Standard Pol Curves

Direct Coated

Control

Very Dry

V. Dry

Technical Accomplishments:DC Cathode Electrochemical Diagnostics

• Current interrupt resistance and high frequency resistance correlate well with performance data in FA conditions, where ohmic losses are very significant

• Electrode ionic resistance is inversely correlated to performance • Direct coating changes water management – modeling has potential to accelerate development

Very Dry

Current Interrupt, High Frequency, & Electrode Ionic Resistance Data

0

50

100

150

200

250

300

CIR (HS, i800)

HFR (HS, 0.2V)

CIR (FA, i400)

HFR (FA, 0.2V)

EIR*10 (HS, 0.2V)

EIR/10 (FA, 0.2V)

Res

ista

nce

[ mΩ

/cm

2 ] Ctl Cathode 0.40 Pt 12 μm GSMDC Cathode 0.40 Pt 12 μm GSM

Wet Very Dry V. DryWet

• Standardized protocol that combines BOL robustness testing with key cathode diagnostics at wet and dry conditions

• Test summary– Pre-Conditioning Diagnostics

• Cleaning Cyclic Voltammograms (CVs)• CV, H2 Cross-Over, Electrochemical Impedance Spectroscopy (EIS)

– Conditioning– Saturated and Super-Saturated Performance

• Polarization Curves, Current Interrupt Resistance, and Stoich Sensitivity– Saturated Diagnostics

• He/O2, O2 Tafel• CV, H2 Cross-Over, EIS

– Sub-Saturated and Hot Sub-Saturated Performance • Polarization Curves, Current Interrupt Resistance, and Stoich Sensitivity

– Sub-Saturated Diagnostics• He/O2, O2 Tafel• CV, H2 Cross-Over, EIS

Technical Accomplishments:DC Cathode Electrochemical Diagnostics

Investigated impact of direct-coated electrode structure

on molecular diffusion

Integrated I-V to quantify oxidized impurities which are associated

with conditioning time

Determined flooding sensitivity is the

biggest gap to close for DC cathode

Technical Accomplishments:Durable 12 μm GORE-SELECT® Membrane

Gore N2 RH Cycling Protocol:

Tcell (C)

Pressure (kPa)

Flow (Anode/Cathode,

cc/min)

80 270 500 N2 / 1000 N2

• Cycle between dry feed gas andhumidified feed gas (sparger bottle temp = 94˚C)

• Dry feed gas hold time: 50 sec.

• Humidified feed gas hold time: 10 sec.

• For further information, reference: W. Liu, M. Crum ECS Transactions 3, 531-540 (2007)

GORE, GORE-SELECT and designs are trademarks of W. L. Gore & Associates, Inc.Note: 12 μm Membrane Testing Not Funded by DOE

0.0E+00

1.0E-08

2.0E-08

3.0E-08

4.0E-08

5.0E-08

6.0E-08

7.0E-08

8.0E-08

9.0E-08

0 2000 4000 6000 8000 10000Life (hours)

170

Pressure(kPa)

50

Inlet RH(%)

60-12010-1.720-100080

Exit RH(%)

Stoic (A and C)

Load(mA/ cm²)T Cell (°C)

170

Pressure(kPa)

50

Inlet RH(%)

60-12010-1.720-100080

Exit RH(%)

Stoic (A and C)

Load(mA/ cm²)T Cell (°C)

Fluo

ride

Rel

ease

Rat

e

• Both MEAs were stopped at 9,000 hrs and were far from failure• H2 crossover at 9000 hr: ≤ 0.017 cc/min.cm2

• Very little change in membrane thickness after 9,000 hrs on test• This test data gathered in 2007 using a membrane that is equivalent

to the 18 µm membrane used in the baseline MEA construction

GORE, GORE-SELECT and designs

are trademarks of W. L. Gore &

Associates, Inc.

Technical Accomplishments:9,000 Hour GORE-SELECT® Membrane Durability in 808C Duty Cycle

Overstress fromTensile tests

Determining Parameters for the Numerical Model

nA&nA )(σε =

( ) )( hvP

v ffKK

Kf ε=⇒+

=Relaxation ExperimentsAt various hold strains

Long-term modulus variation

Log-fit

PK

BA heq += )ln(εσ

Technical Accomplishments: Mechanical Modeling

Visco-Elastic RelaxationSimulation Results Compared with the Experimental Data

Hold strain = 0.5

Hold strain = 0.2

Hold strain = 0.1

Hold strain = 0.05

Equilibrium stress

Uniform displacement loading

Membrane

Technical Accomplishments: Mechanical Modeling

2D Plane Strain Finite Element Model for a Single CellTechnical Accomplishments: Mechanical Modeling

• Membrane’s material properties at various temperature and humidity values acquired through experiments

• Loading consists of pre-stressing and hygrothermal variations

Proposed Future Work for FY10: Summary

Continuedevelopment of

direct coatedelectrodes

Mechanicalmodeling

Optimizereinforced MEA

structure

Fuel cellheat / water

managementmodeling &validation

Optimizeelectrode and

GDL properties

FY11/12 Conditioning Cost review Scale-up Stack Validation

FY 10

FY 11/12

• Determine critical GDL characteristics, understand cost-performance relationships, and select alternative GDL options for fuel cell testing

• Measure unknown critical properties of GDLs: Liquid Water distribution (neutron imaging) and polarization measurement including quantification of mass, ionic, and kinetic losses as well as overall water transport coefficient

• Comprehensive modeling of effects of various GDL, electrode, and membrane parameters on performance, RH distribution, temperature gradient and water accumulation

• Characterize interactions between alternative GDL’s and break-in and performance testing

Proposed Future Work for FY10:Water Management Modeling

Computational Modeling

Segmented Cell

Neutron imaging

SEM

Proposed Future WorkMechanical Modeling

• Static and time dependent (viscoelastic/viscoplastic) testing of baseline materials • Numerical simulation of stress evolution around MEA imperfection

Experimental Data

Cast membrane

Visco-elastic-plastic Properties

Sorption Behavior(λ vs. RH)

Current Work

FY 10

Numerical Modeling

Fracture-FailureModels

Cast membraneMEA

Time Effecton Swelling

Stress-DiffusionModel

e-PTFE reinforced membrane

Fracture and Failures

Flaws: Geometries and locations

NAFION is a registered trademark of E. I. DuPont de Nemours & Company

Proposed Future Work for FY10:

• Coating equipment which had long lead times was purchased in Phase 1• Only 20% of the project budget was spent as of 4/8/10• When direct coating equipment was installed and qualified in March, more

resources were added to the project• Spending and technical progress are both in agreement with the project forecast

Collaborations

• University of Delaware – MEA Mechanical Modeling– A. Karlsson & M. Santare

• Penn State University* – Fuel Cell Heat and Water Management Modeling and Validation– M. Mench

• UTC Power, Inc. – Stack Testing– T. Madden

• W. L. Gore & Associates, Inc. – Project Lead– F. Busby

*New partner added for FY10

Summary (1)• The overall objective of this project is to develop unique, high-

volume manufacturing processes for low-cost, durable, high-power density 3L MEAs that require little or no stack conditioning.

• Approach:–Reduce MEA & Stack Costs

• Reduce the cost of intermediate backer materials • Reduce number & cost of coating passes • Improve safety & reduce process cost by minimizing solvent use • Reduce required conditioning time & costs

–Optimize Durability • Balance tradeoffs between mechanical properties and performance of the 3L

construction

–Unique Enabling Technologies • Direct Coating: Use to form at least one membrane–electrode interface• Gore’s Advanced ePTFE membrane reinforcement & advanced PFSA

ionomers enable durable, high-performance MEAs• Utilize modeling of mechanical stress and heat / water management to

accelerate low-cost MEA optimization• Advanced fuel cell testing & diagnostics

• Key Accomplishments–Cost model results indicate that a new 3L MEA process has potential to

reduce 3L MEA cost by 25%–Development of direct coated anode and cathode electrodes is underway

• Primary and alternative paths have been determined• Current density of direct coated electrodes on reinforced 12 µm

membrane is equivalent to current commercial MEA• Gore has demonstrated 9,000 hour membrane durability

(DOE 2015 target is 5,000 hours)–Development of a layered structure MEA mechanical model using

non-linear (viscoelastic & viscoplastic) polymer and electrode properties which will predict MEA durability for a variety of temperature & relative humidity cycling scenarios is underway

• The combination of Gore’s advanced materials, expertise in MEA manufacturing, & fuel cell testing with the mechanical modeling experience of University of Delaware and the heat and water management experience of Penn State University enables a robust approach to development of a new low-cost MEA manufacturing process

Summary (2)

W. L. Gore & Associates, Inc.• Will Johnson• Glenn Shealy• Mark Edmundson• Wen Liu• Simon Cleghorn• Laura Keough

Department of Energy• Jesse Adams• Jill Gruber• Pete Devlin

Acknowledgements:

University of Delaware• Anette Karlsson• Mike Santare• Melissa Lugo• Narinder Singh

Penn State University• Matthew M. Mench

UTC Power, Inc.• Thomas Madden